- Note that this adds an extra check of the response, which is that
the issuer string in the response must match issuer of the requested
URL.
- Some of the error messages also changed to match the errors provided
by oidc.NewProvider
See RFC6648 which asks that people stop using `X-` on header names.
Also Matt preferred not mentioning "IDP" in the header name.
Signed-off-by: Matt Moyer <moyerm@vmware.com>
This change makes it easier to understand misconfigurations caused
by issuers with extraneous trailing slashes.
Signed-off-by: Mo Khan <mok@vmware.com>
The admin kubeconfigs we have on EKS clusters are a bit different from others, because there is no certificate/key (EKS does not use certificate auth).
This code didn't quite work correctly in that case. The fix is to allow the case where `tlsConfig.GetClientCertificate` is non-nil, but returns a value with no certificates.
Signed-off-by: Matt Moyer <moyerm@vmware.com>
to avoid garbage collection breaking the refresh flow
Also changed the access token lifetime to be 2 minutes instead of 15
since we now have cert caching.
- Use `nickname` claim as an example, which means we only need the `openid` scope.
This is also more stable since emails can change over time.
- Put the OIDCIdentityProvider and Secret into one YAML blob, since they will likely be copy-pasted together anyway.
- Add a separate section for using alternate claims.
- Add a separate section for using a private GitLab instance.
Signed-off-by: Matt Moyer <moyerm@vmware.com>
Some minor edits I came across while reviewing this:
- Capitalize "GitLab" the way they do.
- Use `{{< ref "xyz" >}}` references when linking internally. The advantage of these is that they're "type checked" by Hugo when the site is rendered, so we'll know if we ever break one.
- Add links to the GitLab docs about creating an OAuth client. These also cover adding a group-level or instance-wide application.
- Re-wrap the YAML lines to fit a bit more naturally.
- Add a `namespace` to the YAML examples, so they're more likely to work without tweaks.
- Use "gitlab" instead of "my-oidc-identity-provider" as the example name, for clarity.
- Re-word a few small bits. These are 100% subjective but hopefully an improvement?
Signed-off-by: Matt Moyer <moyerm@vmware.com>
The supervisor treats all events the same hence it must use a
singleton queue.
Updated the integration test to remove the data race caused by
calling methods on testing.T outside of the main test go routine.
Signed-off-by: Monis Khan <mok@vmware.com>
Followup on the previous comment to split apart the ServiceAccount of the kube-cert-agent and the main concierge pod. This is a bit cleaner and ensures that in testing our main Concierge pod never requires any privileged permissions.
Signed-off-by: Matt Moyer <moyerm@vmware.com>
Since 0dfb3e95c5, we no longer directly create the kube-cert-agent Pod, so our "use"
permission on PodSecurityPolicies no longer has the intended effect. Since the deployments controller is now the
one creating pods for us, we need to get the permission on the PodSpec of the target pod instead, which we do somewhat
simply by using the same service account as the main Concierge pods.
We still set `automountServiceAccountToken: false`, so this should not actually give any useful permissions to the
agent pod when running.
Signed-off-by: Matt Moyer <moyerm@vmware.com>
